®

2016 ����*+�����*6��*/01-����-������!�"#$ ��-����

753

LONG-TERM CREEP RUPTURE STRENGTH PREDICITION AND FAILURE MECHANISMS

OF WELDED JOINTS OF BOILER TUBES AND PIPES

KUBO� Zden�k

Material and Metallurgical Research, Ostrava, Czech Republic, EU, creep.lab@mmvyzkum.cz

Abstract

Operating life and reliability of steam boiler and piping systems are considerably determined by long-term

creep rupture strength (CRS) of welded joints of boiler tubes and pipes. Consequently, the appreciable

research effort has been focused to reveal the most important factors affecting the creep properties of

weldments. The presented work deals with the evaluation of creep rupture strength of welded joints in steam

boiler piping system made from low-alloy and chromium modified steels. Cross-weld creep rupture tests and,

at the same time, comparative creep tests of parent material are considered as the optimum approach how to

evaluate the decreasing of creep resistance of weldment towards the base metal and how to quantify this

effect e.g. by using strength reduction factor (SRF) and lifetime reduction factor (LRF) of welded joints. The

necessary condition for such an approach is that the mechanism of creep damage (type IV cracking) in

laboratory tests must be the same that acts in in-plant welded joints of the mentioned steels. The obtained

results confirmed that the precipitation strengthened steels (mainly Cr-Mo-V type) show more pronounced

decreasing of creep rupture strength of the critical locality of welded joint - intercritical part of heat affected

zone when compared to Cr-Mo steels and that they are, therefore, much more prone to the premature failure.

Keywords: Creep, Cr-Mo-V steel, steel P91, strength reduction factor, type IV damage

1. INTRODUCTION

Components in power plants and petrochemical systems operate at high temperatures for long times. A large

proportion of steel pressurised components are manufactured using low alloy ferritic steels and chromium

modified steels, which are selected for their good combination of high temperature creep resistance and cost.

Welded steel pressurised components in particular can experience creep deformation and fracture at high

temperature, and the majority of the creep failures of pressurised components are associated with the welds

[1]. Therefore, the appreciable attention has been focused on revealing the nature of weldment behaviour

during creep exposure and on finding the way how to test them and how to predict the long-term creep strength

of weldments in relation to parent material.

2. WELDMENTS AND TYPE-IV CRACKING

It is well-known that the welding leads to the formation of relatively narrow locality in weld joint with minimum

creep strength [1-3] that

• is surrounded by weld parts with higher creep rupture strength,

• appears as preferable creep failure zone during in-service loading of welded tubes by internal pressure

especially in superposition with additional axial tensile stress,

• limits the economical life of steam boiler piping system.

The critical locality of weldments in heat resistant carbon, low-alloy and chromium modified steels is in most

cases the low-temperature (intercritical) part of heat affected zone (IC HAZ), Figure 1. Its existence then leads

to premature failure designated as type-IV creep failure. This failure type is situated immediately near the

transition of HAZ to the parent material and is typical for the complex stress state of circumferential welds of

®

2016 ����*+�����*6��*/01-����-������!�"#$ ��-����

754

boiler tubes and pipes exposed to internal pressure combined with the axial stress perpendicular to weld axis

and originated from tensile and/or bending loading [3, 4].

Heavily tempered and partially retransformed

microstructure in this region facilitates the creep

failure in IC HAZ. Creep rupture tests using

cross-weld specimens at low stresses (<100

MPa for low alloy steels) show that the type-IV

creep rupture strength is appreciably less than

the parent steel strength [5, 6]. At higher stresses

(>140 MPa for 0.5Cr-Mo-V steel), the cross-weld

strength and parent steel strengths coincide.

Consequently, tests at the high stress levels are

unlikely to reveal the Type IV failure mode as

cross-weld tests will generally fail in the parent

steel or weld metals.

The damage mechanism consists of nucleation,

growth and coalescence of grain boundary

cavities followed by magistral crack propagation

starting usually close to the first subsurface bead

at external surface of tube/pipe.

Probability of type-IV cracking appearance grows

when its operating service oversteps

approximately 5·104 hours, which is evident e.g.

in cases of welded boiler tubes of low-alloy Cr-

Mo-V steels precipitation strengthened by small

particles of carbides and/or nitrides [1, 4, 7].

Owing to thermal effect of welding cycle, type IV zone is characterized by:

• local decrease of precipitation strengthening in consequence to the particles of minor phase particles,

especially of MX type,

• grain size refining in IC HAZ with partly austenitized microstructure enhancing grain boundary sliding

and Coble diffusion creep.

Relatively „soft“ zone in IC HAZ is deformed practically without constraint effect (i.e. independently on the

adjacent „hard“ zones of weld with higher creep strength) during decisive fraction of creep life. In this zone the

accumulation of creep deformation and continual cavitation ahead of creep crack tip leads to the local initiation

of multiaxial stress state. Creep failure is then controlled by the combination of multiaxial stress in superposition

with the maximum principal stress that simultaneously influence both the grain boundary sliding and cavitation

damage accompanied by crack growth [5, 8]. This failure mechanism is typical for weld joints of boiler tubes

from low-alloy vanadium bearing or vanadium free Cr-Mo(V) steels and chromium modified steels of

martensitic type In these cases, the typical feature of preferential type-IV cracking is the pronounced decrease

of stress sensitivity of time to rupture occurring in the region of relatively low stresses and creep lives over 104

hours [1, 3-5, 8-10]. The great attention shall be also focused to the influence of testing condition including the

size factor of test bars [5, 7, 11]. It was found that it is necessary to keep the minimum ratio of cross section

of test bar compared to the width of the weakest weld locality, typically IC HAZ. The narrower width of IC HAZ

(x) and the higher test bar diameter (d), the more pronounced stress redistribution occurs in the course of

creep. Identical failure mode was found in circumferential welds of boiler tubes during in-plant service and

laboratory cross-weld tensile creep tested bars if the x/d ≥ 2 [12, 13].

Figure 1 Schematic drawing of weldment parts

®

2016 ����*+�����*6��*/01-����-������!�"#$ ��-����

755

3. ESTIMATION OF CREEP RUPTURE STRENGTH DECREASE IN WELDS OF TUBES/PIPES

The effect of welding on creep strength of steel can be expressed in the form of weld restriction coefficient Wr.

This parameter should be also utilized at the calculation of minimum design thickness of tubes Sv:

�Ñ �h@O

�.æH��hBI (1)

where p = design steam pressure [MPa]

Di = internal tube diameter [mm]

�D = allowable stress at design temperature T [MPa].

where tr = creep life (time to rupture) [h].

Wrmax is the maximum value of Wr, and is defined for full loading by applied stress in perpendicular direction

to weld axis.

The estimate of Wrmax parameter is based on the results of cross weld creep rupture tests. This procedure

enables to evaluate temperature as well as time to rupture dependence of both creep rupture strength values

RmT(W) and the weakest (critical) weld locality. The more desirable (but also expensive) way how to express

Wrmax is to perform cross weld creep tests simultaneously with creep tests of the base material used for weld

joint preparation and to compare the creep rupture strength of weldment RmT (W) and creep rupture strength

of base metal RmT (BM):

md��� ���

l�J�B

l�JN�KL q�½d � ^ M � (2)

where SRF represents the strength reduction factor [3] as a quite realistic measure of creep rupture strength

decrease of weld in comparison to the respective parent material.

As quite justified we can consider introducing of Wr parameter in eq. (1) in cases of circumferential welds of

boiler tubes characterized by significant increase of axial tensile stress during long-term creep exposure. An

additional axial loading originated in self-weight of tube bundle and/or bending moment action cannot be

avoided, too. In such cases we can expect substantial decrease of long-term creep rupture strength and

shortening of creep rupture life of welds in accordance with experimentally evaluated SRF parameters.

4. EVALUATION OF WELD RESTRICTION PARAMETERS OF TUBES/PIPES MADE OF VARIOUS

CREEP RESISTANT STEELS

Four types of Cr-Mo and/or Cr-Mo-V creep resistant steels representing crucial boiler tube/pipe steel grades

were chosen to demonstrate the long-term behaviour of weld joints and the significance of strength and life

reduction factors. They were:

• low-alloy 0.3 % Mo bearing steel (16Mo3 according to EN 10 216-2),

• low-alloy 2.25 % Cr - 1 % Mo steel (10CrMo9-10 according to EN 10 216-2),

• low-alloy Cr-Mo-V steel grade corresponding to 14MoV6-3 steel according to EN 10 216-2,

• martensitic grade P 91 steel (X10CrMoVNbN9-1 according to EN 10216-2).

Detailed information concerning the fundamental parameters of experimental material, welding technologies,

filler material and PWHT regimes, mechanical properties of weld joints and particularly the results of creep

rupture tests were stated elsewhere [11, 14, 15]. The methods of preparing the welds comprised TIG

technology in orbital automated or manual version and manual arc welding (MMAW) technology. The long-

term creep rupture strength of both weld joints (cross weld stress rupture tests) and parent material used was

investigated on creep test bars of 3.2 to 10 mm in dependence of actual tube/pipe thickness. Test temperatures

and applied stresses were chosen with respect to standardized creep rupture strengths of individual boiler

material and to necessity of extrapolation of creep rupture strength at time over 104 hours. The usual

1≤=≤ ),(max

rrr tTfWW

®

2016 ����*+�����*6��*/01-����-������!�"#$ ��-����

756

mathematical statistics treatment of creep rupture data was applied to describe stress-temperature

dependence of time to rupture including the long-term creep rupture strength parameters evaluation. For these

purposes, Seifert parametric equation [16] was chosen in which the applied stress (or rupture strength RmT) is

expressed by quadratic function of the parameter P in the form:

678��� p� � p� h µ � p. h µ. (3)

µ ^�, � 678½d h ���S (4)

where T is the absolute temperature (K), tr is time to rupture (h) and C, A0 to A3 are optimized constants.

The fundamental characteristics of all examined welds are presented in Table 1 together with the results of

creep rupture strength and also SRF evaluation. The example of time-temperature dependence of creep

rupture strength of weld and corresponding boiler tubes/pipes altogether with resulting SRF values are shown

also in Figure 2 for the weld made of steel 14MoV6-3.

Table 1 Specification of boiler tubes/pipes and welding, results of evaluation of CRS and SRF

Boiler tube/pipe Welding

process Filler material Temp. (°C)

RmT (MPa) SRF

Steel grade Dimension 104 h 105 h 104 h 105 h

16 Mo3 ø51 x 5 mm 141

(TIG)

Union I Mo

(ø 0.8 mm)

500 152 82 0.84 0.77

525 86 34 0.78 0.71

10CrMo9-10 ø320 x 20 mm 111

(MMAW)

OK 76.28

(ø 2.5 to 4 mm)

550 94 61 0.94 0.95

575 69 41 0.91 0.92

600 49 27 0.86 0.87

14MoV6-3 ø 38 x 5 mm 141

(TIG)

C-321

(ø 0.8 mm)

550 120 65 0.92 0.75

575 80 40 0.76 0.65

P91 ø 324 x 32 mm 111

(MMAW)

Fox 9CMV

(ø 3.25 to 4 mm)

550 150 111 0.87 0.79

575 113 81 0.82 0.75

600 83 58 0.77 0.70

625 60 40 0.73 0.66

650 42 27 0.69 0.62

Figure 2 Temperature-time dependence of stress (left) and SRF (right) of 14MoV6-3 steel weldment

It is evident that regardless of the applied welding technology the highest SRF values, among tested steel

grades, has low-alloy Cr-Mo steel. Either Mo-bearing grade 16Mo3 or Cr-Mo-V steel 14MoV6-3 show stronger

®

2016 ����*+�����*6��*/01-����-������!�"#$ ��-����

757

decrease of creep rupture strength of weldments compared to parent material. It also seems that increasing

vanadium content in the steel enhances creep rupture strength but, on the other hand, also decreases SRF

factor in comparison to Cr-Mo steels. Although chromium modified steels have higher creep rupture strength

than above mentioned steels, this is counterweighted by more pronounced decrease of SRF being only from

0.6 to 0.7 for time to rupture 105 hours at 600 to 650 °C. The increase of both temperature and time to rupture

is then accompanied by decrease of SRF in all cases of the examined weld joints.

The typical feature of all boiler tube welds is the

preferable type IV mechanism of creep failure

indicating that the IC HAZ is the weakest weld locality

from the viewpoint of its long-term rupture strength

(Figure 3). This fact corresponds very well with the

observed decrease of stress sensitivity of time to

rupture that is more intensive in weldments compared

to parent metal. The decreasing tendency of time-

temperature dependence of SRF is in many cases

evident already at times to rupture above 103 hours

Type IV cracking (as the principal failure mode of

creep specimens in our experiment) is fully

comparable with damage mode and operation life

exhaustion of in-plant welded boiler tubes/pipes. This

confirms the suitability of the above described approach to the testing of creep properties of welds. Finally, the

minimum diameters of creep specimens were 3.2 mm and, simultaneously, the width of the IC HAZ was lower

than 1 mm, i.e. in all examined cases the required geometrical condition d/x<2 was fulfilled. Type IV cracking

occurs significantly before the expected creep life of the parent steel especially when there is significant tensile

stress in axial direction originated either in self-weight of tube bundle or in additional bending load, when

maximum design allowable stress did not take into account stress restriction in the weldments, when the creep

rupture properties are low due to substandard metallurgical quality, when circumferential welds of boiler

tubes/pipes were made with applying improper welding fillers, heat input and post weld heat treatment regime

and also when the component is in long-term operation at temperatures over 580 °C. Such a behaviour is

typical mainly for strongly precipitation strengthened low alloy and martensitic steels when thermal cycle during

welding causes partial dissolution of dispersed particles of secondary (vanadium rich) MX phase and

tempering of martensite. These processes significantly reduce precipitation strengthening especially in IC HAZ

and tend to the explicit action of type IV cracking controlling the rupture life of weld joints [3, 15, 17].

5. CONCLUSION

Creep life prediction of piping systems in power plants can be evaluated by direct assessment of creep

properties of boiler tubes and their welded joints. Such an attitude gives very precise and reliable results

provided that weld joint is the critical part of the piping system due to the existence of IC HAZ. The size of test

specimens can be neglected, if these diameter is at least twofold compared to the thickness of the IC HAZ. In

this case the same failure mode appears in both creep tests as well as in-service parts. Creep rupture strength

reduction in welds can be then expressed by means of SRF factor.

ACKNOWLEDGEMENTS

This paper was created in the frame of the Project No. LO1203 "Regional Materials Science and

Technology Centre - Feasibility Program" funded by Ministry of Education, Youth and Sports of the

Czech Republic.

Figure 3 Time dependence of SRF of the tested

steels

®

2016 ����*+�����*6��*/01-����-������!�"#$ ��-����

758

REFERENCES

[1] KIMMINS, S. T., COLEMAN, M. C., SMITH, D. J. An overview of creep failure associated with heat affected zone

of ferritic weldments. In: Creep and Fracture of Engineering. Materials and Structures. London: The Institute of

Materials, 1993, pp. 681-694.

[2] PRICE, A. T.-WILLIAMS, J. A. Influence of Welding on Creep Properties of Steels. In: Recent Advances in Creep

and Fracture of Engineering Materials and Structures. Swansea: Pineridge Press, 1982, pp. 265-353.

[3] ETIENNE, C. F., HEERINGS, J. H. Evaluation of the influence of welding on creep resistance. Steel Research,

1994, vol. 65, no. 5, pp. 187-196.

[4] BREAR, J. M., MIDDLETON, C. J. The applicability of weld failure prediction maps to new and exposed welds. In

3rd International HIDA and Integrity Conference, Lisbon: ISQ-Oeiras, 2002, pp. 401-412.

[5] SMITH, D.J., WALKER, N.S., KIMMINS, S.T. Type IV creep cavity accumulation and failure in steel welds. Int. J.

Pres. Ves. and Piping, 2003, vol. 80, no. 9, pp. 617-627.

[6] FRANCIS, J. A., MAZUR, W., BHADESHIA, H. K. D. H. Type IV cracking in ferritic power plant steels. Materials

Science and Technology, 2006, vol. 22, no. 12, pp. 1387-1395.

[7] HYDE, T. H., SUN, W., BECKER, A. A., WILLIAMS, J. A. Life prediction of repaired welds in a pressurised CrMoV

pipe with incorporation of initial damage. Int. J. Pres. Ves. Piping, 2004, vol. 81, no. 1, pp. 1-12.

[8] ALBERT, S. K., et. al. Variation in the type IV cracking behaviour of a high Cr steel weld with post weld heat

treatment. Int. J. Pres. Ves. Piping, 2003, vol. 80, no. 6, pp. 405-413.

[9] HYDE, T. H., BECKER, A. A., SUN, W.-WILLIAMS, J. A. Finite-element creep damage analyses of P91 pipes. Int.

J. Pres. Ves. Piping, 2006, vol. 83, pp. 853-863.

[10] YAGUCHI, M.-MATSUMURA, T.-HOSHINO, K. Evaluation of Long-Term Creep Strength of Welded Joints of

ASME Grades 91, 92 and 122 Type Steels. In ASME Pressure Vessels and Piping Conference: Volume 6:

Materials and Fabrication, Parts A and B. New York: ASME, 2012, pp. 317-326.

[11] SOBOTKA, J. et al Influence of vanadium on creep rupture SRF of boiler tubes. In: Mathematical Modelling of

Weld Phenomena 5, London: IOM Communications, 2001, pp. 479-494

[12] STORESUND, J., TU, S. T. Geometrical effect on creep in cross weld specimens. Int. J. Pres. Ves. Piping, 1995,

vol. 62, no. 2, pp. 179-193.

[13] PARKER, J. D., STRATFORD, G.C. Strain localization in creep testing of samples with heterogeneous

microstructures. Int. J. Pres. Ves. and Piping, 1996, vol. 68, no. 2, pp. 135-143.

[14] SOBOTKA, J et al. Evaluation of creep resistance of weldments with respect to remnant life prediction of power

equipment (Hodnocení žárupevnosti svar2 ve vztahu k predikci životnosti energetických za�ízení). In Zváranie v

energetike. Trnava: ESAB Slovakia-MtF STU Trnava, 2004, pp. 62-81 (in Czech).

[15] KUBO�, Z., SOBOTKA, J. Assessment of long-term creep properties, creep failure and SRF of boiler tube

weldments. In. BALTICA VI-Life Management and Maintenance for Power Plants, Helsinki: VTT Industrial

Systems, 2004, pp. 477-490.

[16] SEIFERT, W. Statistische Kenngrössen aus der Ausvertung von Zeitstandversuchen. In. Warmfeste Metallische

Werkstoffe 6, Zittau: Kammer der Technik, 1987, pp. 129-135.

[17] PARKER, J. Factors affecting Type IV creep damage in Grade 91 steel welds. Materials Science and Engineering

A, 2013, vol. 578, pp. 430-437.